Bombesin, platelet-derived growth factor, and diacylglycerol induce selective membrane association and down-regulation of protein kinase C isotypes in Swiss 3T3 cells.

Swiss 3T3 cells contain protein kinase C (PKC) isotypes alpha, delta, epsilon and zeta (Olivier, A. R., and Parker, P. J. (1992) J. Cell. Physiol. 152, 240-244). Acute stimulation of quiescent cells with the neuropeptide bombesin decreases the mobility of PKC-delta and PKC-epsilon on SDS-polyacrylamide gels. These slower migrating forms of PKC-delta and PKC-epsilon rapidly (within 1 s) and selectively are found associated with the Triton X-100-soluble membrane fraction. No change in the mobility or distribution of PKC-alpha or PKC-zeta is detected. Long-term treatment of cells with bombesin induces selective membrane association and down-regulation of PKC-delta and PKC-epsilon (decreasing 70 and 65%, respectively). No change in the long-term distribution of PKC-alpha and PKC-zeta was detected. Bombesin did, however, increase PKC-alpha protein levels by 60% compared to control cells. PKC-zeta levels remained unchanged. Both the shift in mobility and down-regulation of PKC-delta and PKC-epsilon were only induced by mitogenic doses of bombesin. The potent mitogen platelet-derived growth factor induced similar effects on the PKC isotypes delta and epsilon. PKC-alpha and PKC-zeta levels were unaffected. Repeated doses of the synthetic diglyceride 1-oleoyl-2-acetyl-sn-glycerol induced PKC-delta and PKC-epsilon down-regulation and stimulated the cells to divide. Again PKC-alpha and PKC-zeta levels were unaffected. These results show a correlation between the membrane association and down-regulation of PKC-delta and PKC-epsilon and the entry of cells into S phase.

* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. $ To whom correspondence and reprint requests should be addressed: Protein Phosphorylation Laboratory, Imperial Cancer Research Fund, P. 0. Box 123, Lincoln's Inn Fields, London WC2A 3PX, United Kingdom.
The events leading to cell proliferation have been extensively studied using Swiss 3T3 cells as a model system (5). Recently it was reported that fibroblasts express PKC-6, -e, and -i in addition to PKC-a (6-8). These enzymes are differentially downregulated in response to phorbol esters, implying different regulatiodcompartmentalization for each isotype (6,7). While phorbol esters provide pharmacological evidence of distinct operation of these PKC isotypes, their rolehehavior in response to physiological agonists with defined mitogenic properties has not been investigated previously in fibroblasts. In view of the complex, multiphasic nature of DAG production in response to various cellular agonists (e.g. Refs. [9][10][11] it has been pertinent to determine directly how such agonists impinge upon specific PKC isotypes. Bombesin binds to a G-protein-coupled receptor (12), inducing a number of early responses including phospholipid hydrolysis (5,(13)(14)(15) which results in the acute changes in second messengers such as Ca2+ and DAG that could activate PKC (13,14). Acting through its tyrosine kinase receptor, plateletderived growth factor (PDGF) can also induce a similar series of events to generate second messengers (16), including those involved in PKC activation. As with other mitogens, stimulation needs to be sustained for a prolonged period in order to commit cells to division. The prolonged signals that are required to establish commitment are poorly understood.
With the use of isotype-specific antisera it is shown here that both bombesin and PDGF can induce marked changes in PKC isotype expression and distribution after both acute and prolonged exposure to the mitogens. In addition, the repeated addition of DAG can mimic these prolonged effects. The results are consistent with a role for PKC activation and consequent down-regulation in GI progression, perhaps contributing to the sustained mitogen responses necessary for commitment to S phase.
Cell Culture and Extraction-Swiss 3T3 cells were maintained as previously described (6). Cells were judged quiescent after 8 days, a time at which 9597% of the cells were in G,-G, as judged by flow cytometry (see Fig. 8) and immunofluorescence (data not shown). Cells The immunoblots were visualized using the ECL detection reagents and autoradiography as described under "Materials and Methods." This is one of five similar dose-responses.
were stimulated with growth factors or lipids for increasing times as indicated in the text or figure legends. For total cell lysates, the cells were placed on ice and washed two times with phosphate-buffered saline (PBS). %ce concentrated Laemmli (21) sample buffer (150 pl/6-cm plate, 300 gl0-em plate) was added to each plate. The cell lysate was scraped to the side of the plate and transferred to a tube which was heated to 95 "C for 10 min. The lysate was vortexed and centrifuged for 5 min in a microcentrifuge at maximum speed (14,000 rprn). An aliquot from the supernatant was taken for protein determination (22). The remaining portion of the sample was frozen and stored in liquid Nz until use.
For cell fractionation (translocation) experiments, one 15-dish of cells (seeded a t 3 x 106 celldplate) was used per time point. After stimulation the cell layer was washed with Buffer A (25 n" Tris-HCI, pH 7.5.250 m~ sucrose, 2.5 m~ magnesium acetate, 2 n" dithiothreitol, 10 n" benzamidine, 10 n" sodium fluoride). As much of the buffer as possible was removed and replaced with 1 ml of homogenization buffer (Buffer A containing 5 n" EGTA, 5 n" EDTA plus protease inhibitors, 250 pg/ml leupeptin, 150 pg/ml aprotinin, and 1 m~ phenylmethylsulfonyl fluoride). The cells were scraped to the side of the plate and homogenized for 30 strokes in a Dounce homogenizer with a tight fitting pestel. The homogenate was then centrifuged for 10 min at 90.000 rpm in a Beckman TL 100 centrifuge at 4 "C. The supernatant (cytosol) was taken and 500 of 4 x concentrated Laemmli sample buffer was added and the samples heated to 95 "C for 10 min. The pellet was extracted for 20 min on ice in 1 ml of Buffer B containing 25 m~ Tris-HC1, pH 7.5,5 m~ dithiothreitol, 10 n" benzamidine, 10 n" sodium fluoride, 5 m~ EDTA, and 1% Triton X-100 plus protease inhibitors as above. The detergent-soluble proteins were then collected by centrifugation at 4 "C for 20 min at 12,000 rpm. The supernatant containing the detergentsoluble fraction (membranes) was taken and 500 p1 of 4 x Laemmli sample buffer was added and heated as above. The pellet (detergentinsoluble fraction) was resuspended in 1.5 ml of 1.3 x Laemmli sample buffer and heated to 95 "C as above. All samples were frozen and stored in liquid N2.
Phosphatase Deatment of PKC-+Quiescent cells, or cells stimulated with bombesin for 15 min, were lysed in Buffer C containing 80 n" pglycerophosphate, 20 n" MES, pH 6.5,lO n" EGTA, 2 m~ EDTA, 1 n" benzamidine, 20 n" &mercaptoethanol, 1% Triton X-100 plus the protease inhibitors described above. The supernatant from a 12,000 x g centrifugation for 15 min was loaded onto a Fast Flow S column equilibrated in Buffer D (20 m~ MES, pH 6.5,0.5 n" EGTA, 0.5 m~ EDTA, 20 n" pmercaptoethanol, 1 m~ benzamidine plus 20 m~ NaCl and protease inhibitors). After the column was washed with 10 x volume of Buffer D, the bound proteins were eluted with Buffer D containing 400 n" NaCl. Western blot analysis showed that only PKC-e bound to the column, while PKC-6 and PKC-a were recovered in the flow through (data not shown). The eluate was divided into aliquots and each treated with 5 uniWml protein phosphatase 2A catalytic subunit diluted in Buffer E (20 m~ Tris, pH 7.5, 0.5 m~ dithiothreitol, 1 mdml bovine serum albumin) at 30 O C for 30 min either in the presence or absence of 1 p~ microcystin. The reaction was stopped by the addition of 2 x Laemmli sample buffer, boiled, and subjected to SDS-PAGE. Zmmunoblot A m l y s S a m p l e s containing equal amounts of protein (30-50 pg), or derived from equivalent numbers of cells (translocation studies), were loaded in each lane for SDS-polyacrylamide gel electrophoresis (SDSPAGE) (21). The bisacrylamide concentration was reduced to 0.075% to resolve the phosphorylation states of PKC-E. This, however, had no effect on the resolution of the other PKC isotypes. The gels were transferred and analyzed by immunoblot analysis as previously described (6). To reprobe the membranes with a different antibody they were wet and soaked in 100 n" glycine, pH 2.5, for 10 min at room temperature. The solution was removed and the blots neutralized in 100 m~ Tris, pH 7.5, washed in PBS, and blocked again in PBS containing 5% milk powder and probed as previously described (6).
Flow Cytometric Analysis-After stimulation of the cells (6 cm dish) the medium was gently removed and the cells detached from the plate by trypsinEDTA treatment. The cells were transferred to a tube and

Effect of Acute Bombesin Deatnent on PKC Zsotypes-
Purification of recombinant PKC-6 (19,24) or endogenous PKC-6 and PKC-c from tissue (25-27) have shown that these enzymes exist as phosphorylated entities displaying heterogeneity on SDS-PAGE. The neuropeptide bombesin has been shown to induce a number of early events in Swiss 3T3 cells, including increased phosphorylation of proteins (5, 28). To determine whether this agonist has any effect on the mobility of these PKCs on SDS-PAGE, Swiss 3T3 cells were stimulated for 15 min with increasing doses of bombesin. This acute treatment of the quiescent cells is found to induce a size shift of both PKC-6 and PKC-c as seen by analysis on SDS-PAGE and subsequent Western blotting (Fig. 1). The extent of this shift is dependent on the bombesin concentration, maximum shift is achieved by mitogenic concentrations (see below and Ref. 29) of the neuropeptide (Fig. 1, lanes 5-7). PKC-e appears as a very distinct doublet when analyzed on 7.5% SDS-PAGE with a lower bisacrylamide concentration (see Fig. 1). To determine whether this slower migrating form is a result of phosphorylation of this enzyme, extracts from untreated and bombesintreated extracts were subjected to Fast Flow S chromatography to remove the endogenous serine threonine phosphatases (30). The eluates were then incubated with protein phosphatase 2A either in the presence or absence of the phosphatase inhibitor microcystin. The results in Fig. 2 show that treatment of both quiescent and stimulated cell preparations caused an increase in the mobility of the respective proteins and microcystin blocked this effect. Dephosphorylation of PKC-E in quiescent cells resulted in an increase in the mobility of the protein (denoted a in Fig. 2), indicating that PKC-c has a basal level of phosphorylation (denoted b in Fig. 2) in quiescent cells (Fig. 2,  compare lune 3 with I , 4, and 5). Treatment of the stimulated cell eluates with phosphatase caused the dephosphorylation of both of the bands resulting only in the lower molecular weight forms (a and b). These results therefore indicate that PKC-e is a phosphoprotein and stimulation with bombesin increases this phosphorylation. This is most likely also the case for PKC-6 (25, 261, however, unlike PKC-e, it was not possible to separate PKC-6 from endogenous protein phosphatases by cation exchange chromatography (not shown). PKC-a is also known to be phosphorylated in response to growth fadors (31). A shift in mobility is, however, not detected (Fig. 1). PKC-( does not appear to change migration under any condition. The activation of PKC is frequently measured through its ability to associate stably with membranes upon acute treatment with phorbol esters or growth factors that induce DAG production (1, 2). To determine which of the identified PKC isotypes can translocate in response to bombesin, fractionated cell lysates were subjected to 10% SDS-PAGE and the PKC content analyzed by Western blotting. Fig. 3 shows the time dependent distribution of the different PKC isotypes between the cytosolic and detergent-soluble membrane fractions in response to mitogenic concentrations of bombesin. In quiescent cells the PKC isotypes 6, e, and ( are distributed almost equally between the cytosol and membrane fractions. PKC-( is also present in the detergent-insoluble fraction (data not shown). PKC-a, however, is mainly cytosolic. Treatment with bombesin induces rapid association (within 1 s) of PKCd and PKC-e with the membrane fraction and their respective loss from the cytosolic fraction (Fig. 3). In addition to translocation, bombesin induces maximal size shiR of PKC-6 and PKC-e within 1 s of stimulation and it is this form which is associated with the membrane (Fig. 3, and data not shown). The shift in mobility of PKC-e is not as well resolved on this gel system compared to the 7.5% SDS-PAGE low bisacrylamide used in Figs. 1,2,4 therefore, show that bombesin induces an acute shift in apparent size of PKC-6 and PKC-e and that this slower migrating form is predominantly associated with the membrane.
Effect of Prolonged Deatment of Mitogens on PKC Zsotypes-A well documented characteristic of phorbol esters is to induce an increased rate of degradation of PKC when administered for prolonged periods (32). Until recently investigators have been unable to show that down-regulation occurs in vivo upon physiological stimulation. It has only been since additional members of the PKC family were identified that it was found that PKC-e down-regulates in response to prolonged TRH treatment in GH& cells (33-35). To determine whether bombesin has any effect on the stability of the PKC isotypes, 6, e, and (, cells were treated for up to 30 h with increasing doses of the neuropeptide. Fig. 4 shows that mitogenic concentrations of bombesin (12.5-50 nglml) lead to partial down-regulation of PKC-6 and PKC-e (Fig. 4, lanes 5-7), resulting in a 70 and 65% decrease in the PKC-6 and PKC-e levels, respectively, at 30 h post-stimulation (Table I). PKC-a levels, however, appear to be induced upon bombesin treatment increasing 60% above control levels at this time ( Fig. 4 and Table I). The down-regulation of PKC-6 and PKC-e, and induction of PKC-a, are time dependent as shown in Fig. 5B, where cells were treated with bombesin for up to 44 h. Under no conditions did PKC-( levels change. By comparison, a non-mitogenic dose of bombesin had no reproducible effect on the expression levels of the PKC proteins throughout the time course tested (Fig. 5A).
If activation is reflected by the stable association of PKC isotypes with the membrane fraction, the results would suggest that only the two Ca2+-independent isotypes (6 and e) are activated by acute bombesin treatment. However, it was important to determine whether bombesin could also induce membrane association of these different enzymes at later times during progression through the cell cycle. Cell extracts were fractionated after prolonged treatment with bombesin and the PKC content in each fraction determined by Western blotting after electrophoresis on 10% SDS-PAGE. Fig. 6 shows that even a t later times, between 7 and 36 h post-stimulation, the major proportion of PKC-6 and PKC-e is associated with the detergent-soluble membrane fraction. Also, as with the acute treatments, the distribution of PKC-a and PKC-( does not appear to change. The decrease in the PKC-I; levels seen at 36 h is not reproducible. In each case there is a net change in the amount of PKC-6, -e, and -a at the later times, consistent with the data in Figs. 4 and 5 and Table I. The results imply that it is the membrane-associated populations of PKC-6 and -e which are down-regulated. Effect of PDGF on PKC Zsotypes"Ib determine whether PDGF, a major serum growth factor for these cells, would elicit similar effects on the PKC isotypes, cells were treated for increasing times with mitogenic concentrations of PDGF and the PKC content at each time was evaluated by Western blotting. Fig. 7 shows that PDGF also induces a size shift of PKC-6 and PKC-e at early times, 15 min and 1 h. Both of these PKC isotypes are also partially down-regulated at later times, where the protein levels decreased by approximately 60% after 19 h for both PKC-6 and PKC-e (Table I). PDGF, however, does not induce PKC-a expression. The slight increase at 24 and 36 h after treatment was not reproducible. PKC-C sometimes appears as a doublet on 8% SDS-PAGE but no change is ever seen in stimulated compared to control cells. These effects of PDGF, like those of bombesin, are dose-dependent and correlate with mitogenicity (data not shown). PDGF also induces membrane association of PKC-6 and PKC-e, but not of PKC-a and PKC-C under the extraction and fractionation conditions described for bombesin (data not shown).
Taken together, the results show that bombesin and PDGF can elicit both acute and prolonged effects on PKC-6 and -e; bombesin also induces PKC-a expression. Neither mitogen had reproducible effects upon PKC-6.
Effect of Exongenously Added Diacylglycerols on PKC Zsotypes-It has been reported that bombesin and PDGF induce biphasic production of DAG in Swiss 3T3 cells (9,10,13,14). However, a single dose of OAG did cause a size shift (still detectable a t 12 h) with both PKC-6 and PKC-e (Fig. 8). DiC8, another DAG, did not cause this prolonged shift, but did cause down-regulation of these isotypes after continuous treatment (data not shown). The addition of insulin had no significant effect on the PKC levels. These results, therefore, suggest that continuous exposure of cells to exogenously added DAG can induce activation and down-regulation of a t least PKC-6 and PKC-e and support the notion that agonist-induced sustained DAG production activates and induces down-regulation of these PKC isotypes in vivo.
Induction of DNA Synthesis-In order to assess the relationship between the stimulation of DNA synthesis and the effects induced by bombesin and OAG on the different PKC isotypes, the DNA content of cells was analyzed by flow cytometry. First, cells were treated with increasing bombesin concentrations for 26 h. Fig. 9A (panels 4 and 5) shows that doses between 25 and 100 ng/ml induce the maximal number of cells to enter S phase (for bombesin this is only approximately 15%). These are similar to the doses at which the shift in mobility and down-regulation of PKC-6 and PKC-e occur (Figs. 1 and 4).
In order to determine whether a single dose or continuous treatment of OAG is sufficient to induce DNA synthesis, cells were treated with diglyceride in the absence or presence of insulin. The results show that only the cells treated continuously with OAG enter S phase (Fig. 9B, panels 9 and 10). Insulin potentiates this effect (Fig. 9B, panel 10). A single dose of OAG either in the absence or presence of insulin is not sufficient to stimulate a significant number of cells to initiate DNA synthesis (Fig. 9B, panels 7 and 8). Again, as with bombesin, the activation and down-regulation of PKC-6 and PKC-E induced by OAG correlate with the ability of the cells to enter S phase ( Fig. 8 and Fig. 9B).

DISCUSSION
The data presented here provide evidence that chronic treatment of cells with specific agonists leads to a late phase translocation of particular PKC isotypes (Fig. 6). Associated with this late phase, prolonged activation, is a loss of steady state expression of these PKC isotypes that emulates a typical downregulation response (Figs. [4][5][6]. The ability of sustained DAG to induce this response is demonstrated directly through the use of membrane permeant DAG species. In the context of this cell proliferation model system, the mitogens bombesin and PDGF elicit the acute and prolonged changes in PKC-6 and PKC-e in a way that parallels their potency in stimulating cell division. The implication is that activation of these particular PKC iso- types is a t least a biphasic process that forms part of the program of events induced by these mitogens leading to cell division. Surprisingly, under these growth conditions, there is no evidence of an acute translocation of PKC-a even in the context of the very rapid translocation of PKC-Gk This is in contrast to phorbol ester treatment which under identical extraction conditions induces membrane association of PKC-a (Ref. 6, and data not shown). Whether this reflects some alteration in Ca2+ homeostasis is not clear, although in view of the specific Ca2+ dependence of PKC-a this may prove to be the case. Elevation of Ca2+ during extraction induces membrane association of PKC-a (6,33). During the revision of the manuscript, Ha and Exton (37) reported that PDGF failed to induce membrane association of PKC-a in IIC9 cells, whereas PKC-E did translocate to the membrane. By contrast the translocation of PKC-a was observed under conditions where Ca2+ was substantially elevated (37,38).
Coincident with acute translocation, mitogenic concentra- tions of bombesin and PDGF also induce an acute shift in apparent size of PKC-6 and PKC-e as determined on SDS-PAGE (Figs. 1, 2, 5, and 6). It appears to be this population which is associated with the membrane (Fig. 3). This decrease in mobility is most probably due to increased phosphorylation of these proteins, since as had been shown for PKC-6 and PKC-E purified from rat brain (25,26), the mobility of PKC-e from Swiss cells has been shown to increase after phosphatase treatment (Fig. 2). It is not clear whether this increased phosphorylation is due to autophosphorylation, the activation of a different kinase, or in fact involves the inhibition of a phosphatase. It is of interest that during late phase translocation (and presumably activation), the apparent hyperphosphorylation of PKC-6 and -e is much less marked. This might imply that part of this induced phosphorylation is due to some other kinase activity.
The selective down-regulation of PKC-6 and PKC-E might be caused by sustained levels of DAG induced by bombesin and PDGF. As yet the latest time determined for DAG production is 2 h post-stimulation in Swiss 3T3 cells (9,10). Preliminary evidence indicates that DAG is elevated a t late times following bombesin treatment of Swiss 3T3 c e h 2 Studies in GH4C1 cells have correlated clearly a prolonged phase of DAG production (up to 12 h) with the down-regulation of PKC-e (35). The results shown here address this problem directly by demonstrating that exposure of cells for prolonged periods to the synthetic diglycerides, OAG or DiC8, induces PKC-6 and PKC-E downregulation (Fig. 8, and data not shown). Previous investigators were unable to show down-regulation of PKC-a protein or activity in response to these DAGS (391, probably due to the fact that only a single dose of the diglyceride was employed which is likely to be metabolized very rapidly. This appears to be the case in HG60 cells treated with DiC8, where a single dose did not induce differentiation (40). The results here are consistent with these data since a single dose of OAG or DiC8 (data not shown) fails to induce initiation of S phase in these cells, while repeated additions induce a marked increase in the number of cells that enter into S phase (Fig. 9). Moreover, only the continuous exposure of the cells to diglyceride leads to the selective  down-regulation of PKC-6 and PKC-E, while PKC-a! and PKC-6 are unaffected. Therefore, a good correlation exists between the presumed activation (as judged by translocation) and downregulation of these PKC isotypes and the ability of the cells to enter S phase.
The results support the notion that distinct phases of DAG (9,10) produced in response to Swiss 3T3 cell mitogens lead to the selective activation of PKC-6 and PKC-e. In the prolonged phase of activation during late GI (and beyond) there is a steady decline in the level of PKC-6 and PKC-c expression. Whether the activation of these PKC isotypes or their induced depletion contribute to cell cycle progression, remains to be established.